RE: Onbar - a few funnies
But does anybody know how you set the SHMMAX kernel parameter for SuSE
SLES 8? I have tried
echo aaa > prod/sys/kernel/shmmax
But that doesn't remain after a reboot.
I was told to edit /etc/sysconfig/sysctl but that had no effect.
And I can change SHMALL with PowerTweak, but only SHMALL
Regards
Malcolm
-----Original Message-----
From: owner-informix-list@iiug.org [mailto:owner-informix-list@iiug.org]
On Behalf Of Savio Pinto (s)
Sent: 05 May 2004 22:45
To: informix-list@iiug.org
Subject: RE: Onbar - a few funnies
looks like the onbar process is having problem detaching itself from
the shared memory, may be you have to increase shmmax kernel parameter.
-----Original Message-----
From: owner-informix-list@iiug.org
[mailto:owner-informix-list@iiug.org]On Behalf Of briceavila@hotmail.com
Sent: Wednesday, May 05, 2004 3:32 PM
To: informix-list@iiug.org
Subject: Re: Onbar - a few funnies
There are some kernel parameter "suggestions" in the release notes:
"$INFORMIXDIR/release/en_us/0333/IDS_7.3" or a similar file name. Try
setting these. Hope this helps.
Brice Avila
Minneapolis, Minnesota
malcolm weallans <malcolm.iiug@btopenworld.com> wrote in message
news:<c78df5$k1k$1@terabinaries.xmission.com>...
> --0-12404005-1083683320=:32114
> Content-Type: text/plain; charset=iso-8859-1
> Content-Transfer-Encoding: 8bit
>
> I seem to have got onbar working despite the shmdt:errno 22
messages. But when I looked in bar_act.log I discovered that I was
getting a message telling me that the logical logs were full and I might
have problems with updating sysutils. But when I checked my logs were
nowhere near full. In fact the application isn't running, and the
database is created with no logging. The process continued to backup
just one logical log. I suspect the checks on full logs might not be
working.
> I then turned on BAR_DBUG at level 9 and obtained a complete log of
the whole backup. To my amazement there seems to be no correlation
between the times of the shmdt errors and the events in bar_dbug.log.
The first shmdt error occurred at 12:46:05 but the first lline in
bar_dbug.log is time 12:46:06 and is the onbar -b -L 0 line. It
certainly seems to be the case that the shmdt errors precede begin and
complete backp messages for the dbspaces by about 1 second.
> I continue to investigate.
>
> Any help, as always, is much appreciated.
>
> Malcolm
>
> --0-12404005-1083683320=:32114
> Content-Type: text/html; charset=iso-8859-1
> Content-Transfer-Encoding: 8bit
>
> <DIV>I seem to have got onbar working despite the shmdt:errno 22
messages. But when I looked in bar_act.log I discovered that I was
getting a message telling me that the logical logs were full and I might
have problems with updating sysutils. But when I checked my logs
were nowhere near full. In fact the application isn't running, and
the database is created with no logging. The process continued to
backup just one logical log. I suspect the checks on full logs
might not be working.</DIV>
> <DIV>I then turned on BAR_DBUG at level 9 and obtained a complete
log of the whole backup. To my amazement there seems to be no
correlation between the times of the shmdt errors and the events in
bar_dbug.log. The first shmdt error occurred at 12:46:05 but the
first lline in bar_dbug.log is time 12:46:06 and is the onbar -b -L 0
line. It certainly seems to be the case that the shmdt errors
precede begin and complete backp messages for the dbspaces by about 1
second. </DIV>
> <DIV>I continue to investigate.</DIV>
> <DIV> </DIV>
> <DIV>Any help, as always, is much appreciated.</DIV> <DIV> </DIV>
> <DIV>Malcolm</DIV>
> --0-12404005-1083683320=:32114--
> sending to informix-list
sending to informix-list
sending to informix-list